Digital recording and reproducing apparatuses for video and audio using disk media have been coming into wide use. One of the featured functions of disk media, distinct from tape media, is the function of non-destructive editing, or also called the function of non-linear editing. This function is to provide the capability of reproducing any sections of AV streams in a desired order without actual movement or copy of the AV streams recorded on the disk, and this function is achieved by creating information (playback management information) that indicates the start and end of every section to be reproduced in AV streams and their order of reproduction and implementing reproduction following that information.
In this way, with such disk media, it is possible to make an edit without rewriting source material or moving the data. However, there are some cases where the source material needs to be directly edited. For example, suppose the non-destructive edited result is wanted to be brought together into a single file for easy handling by a personal computer (PC). In this case, only that being used in the edited result should be picked out from associated AV streams and united into a single file.
There is also a case where an intermediate part that is unnecessary in an AV stream is wanted to be deleted in order to increase the empty capacity of the disk. In this case, the parts located before and after the intermediate part should be united.
For either case, plural AV streams should be put together. However, there is a concern that some reproduction noise might occur at the seam when a video data encoding scheme based on the MPEG video standard (ISO/IEC 11172-2 or ISO/IEC 13818-2) is adopted.
The reason is as follows. The MPEG video standard adopts variable length coding, and this specifies that encoding of data to be coded at a predetermined rate is implemented such that a model hypothetical decoder called VBV (Video Buffering Verifier) which should be connected to the output from the encoder will not overflow or underflow.
In this model, coded data is supplied to the VBV buffer at a rate not greater than the aforementioned predetermined rate, and the amount of data occupied in the VBV increases at that rate. On the other hand, the moment one frame or field has been decoded, the occupancy of data decreases instantly by the amount of the corresponding coded data.
Any coded data based on MPEG video cannot be assured to be reproduced correctly if the data has not been encoded in such control that the VBV buffer will not overflow or underflow even if the amount of data repeatedly increases and decreases, as shown in FIG. 22. The risk of reproduction noise when some pieces of video data are joined can be attributed to the possibility of the VBV buffer causing overflow or underflow at a point of the seam.
The reason for the collapse of the VBV buffer at the seam will be described referring to an example. Here, description is made of a case where the front part before time OUT of coded video data A having a time-dependent variation in the occupancy of the VBV buffer shown in FIG. 23 (a) and the rear part after time IN of coded video data B shown in FIG. 23 (b) are joined.
FIG. 23 (c) is the joined result. It is understood that in this case the buffer underflows because a frame or field containing a large amount of coded data takes place right after the point of junction regardless of the low buffer occupancy right before the joined point. The reason why an event of this kind happens is that there is a possibility that the conformity of the buffer occupancy is lost.
In order to solve the above problem, Japanese Patent Application Laid-open Hei 9 No. 182024 proposes a technique of preventing underflow by increasing the transfer speed of the input data to the decoder. However, this method needs a special decoder, resulting in cost disadvantage.
As another method, Japanese Patent Application Laid-open Hei 8 No. 251582 proposes a technique (re-coding) whereby the seam portion upon joining is once decoded and then encoded again so that the amount of the coded data will be kept so as not to cause corruption of the VBV buffer. However, in this case, there is a concern of occurrence of image degradation due to the re-coding process. Further this method needs to implement coding and decoding successively or in parallel, entailing the problem in that the apparatus becomes more complicated.
The present invention has been devised in view of the above problems, it is therefore an object of the present invention to provide a data recording method, a data editing method and a data decoding method as well as a data recorder and a data decoder, which, by a simple configuration, can prevent reproduction noise upon reproduction of an AV stream which is formed of joined AV streams.